Optimization

Elemental contains a variety of routines for solving optimization problems (currently, first-order methods via ADMM) as well as a collection of common utilities, such as Singular Value soft-Thresholding (SVT).

Please note that a large percentage of the following optimizaton routines were inspired by the MATLAB scripts associated with the paper:

S. Boyd, N. Parikh, E. Chu, B. Peleato, and J. Eckstein, Distributed optimization and statistical learning via the Alternating Direction Method of Multipliers,

which describes a wide variety of optimization schemes using the Alternating Direction Method of Multipliers (ADMM).

• Basis pursuit
• Least absolute shrinkage and selection operator
• Linear programs
• Logistic regression
• Model fit
• Non-negative least squares
• Non-negative matrix factorizations
• Quadratic programs
• Robust principal component analysis
• Sparse inverse covariance selection
• Support vector machines
• Utilities